About the missing reconnect issue: Laine noticed about this, but also noticed that the QEMU event we need is current missing, so libvirt can't fix it (yet). Laurent is trying to add it here: https://patchew.org/QEMU/20250217092550.1172055-1-lvivier@redhat.com/. 

Refer to https://issues.redhat.com/browse/RHEL-69455?focusedId=26627330&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-26627330

What were you trying to do that didn't work?

When VM is running, kill the passt process, it will not create a new passt process.

What is the impact of this issue to you?

Please provide the package NVR for which the bug is seen:

libvirt-10.10.0-7.el9.x86_64

How reproducible is this bug?:

100%

Steps to reproduce

1.  start a vm with vhostuser type interface with passt backend

   # virsh dumpxml rhel --xpath //interface
   <interface type="vhostuser">
     <mac address="52:54:00:5a:2d:da"/>
     <model type="virtio"/>
     <backend type="passt" logFile="/run/user/107/passt.log"/>
     <alias name="net0"/>
     <address type="pci" domain="0x0000" bus="0x01" slot="0x00" function="0x0"/>
   </interface>
   

2.  check the passt process, then kill the process

   # ps aux | grep 'passt --vhost-user' | grep -v grep 
   qemu       51879  0.0  0.0 2285416 227
   44 ?       Ss   11:47   0:00 passt --vhost-user --one-off --socket /run/libvirt/qemu/passt/4-rhel-net0.socket --pid /run/libvirt/qemu/passt/4-rhel-net0-passt.pid --log-file /run/user/107/passt.log
   # kill -9 51879
   # ps aux | grep 'passt --vhost-user' | grep -v grep 
   (none)
   # pidof passt
   (none)
   

Actual results

Passt process can not be recovered after killed, which cause the interface down

Expected results

When vm is running, but the passt process killed accidentally, it should be recovered.